Dragonsnakes and Filesnakes Revisited

I've written about both filesnakes (family Acrochordidae) and dragonsnakes (part of the family Xenodermidae¹) before. Traditional snake taxonomy suggests that, although they branch off from the main stem of the snake family tree at about the same time, they're not very closely related. But, new evidence emphasizes the uniqueness of dragonsnakes and thickens the plot in the unfolding story of the evolution of snakes.

Two hypotheses about the relationships of the major groups of snakes. Left: tree based on nuclear genes, showing Acrochordidae and Xenodermidae as successive outgroups to core Colubroidea Right: tree based on mitochondrial genes, showing a sister relationship between Acrochordidae and Xenodermidae From Oguiura et al. 2009

Most phylogenetic analyses are pretty consistent in classifying both filesnakes and dragonsnakes as caenophidians, or "advanced" snakes. But, they differ in their placement of dragonsnakes and other xenodermids, including the truly strange and obscure odd-scaled snakes (Achalinus), bearded snakes (Fimbrios), stream, earth, or red snakes (Stoliczkia²), wood, mountain, or narrow-headed snakes (Xylophis), and a new genus, just described in 2015 and still without a common name, Parafimbrios. Most analyses group xenodermids with the colubroids (pareids¹, vipers, homalopsids, colubrids, lamprophiids, and elapids), albeit as the most basal branch. Many textbooks actually define Caenophidia as Colubroidea + Acrochordidae (aka Acrochordoidea), distinctly separating the colubroids from the filesnakes on the basis of shared, derived characteristics such as wide ventral scales, as well as features of the skull, hemipenes, and the muscles, cartilages, and arteries between the ribs. However, several recent trees based on DNA sequences suggest instead that filesnakes and dragonsnakes might be one another's closest living relatives.

Study	Acrochordid-Xenodermid Relationship	Support	How many species?	What data were used?
Boulenger 1893	X&A	-	-	Morphology
Kraus & Brown 1998	X+A	8%	37	ND4
Kelly et al. 2003	X+A	98%	98	4 mitochondrial genes
Vidal et al. 2007	A,X	not reported	25	7 nuclear genes
Wiens et al. 2008	A,X	>95%	50	20 nuclear genes
Wang et al. 2009	A,X*	100%	30	12 nuclear genes
Zaher et al. 2009	A,X	94%	131	2 mitochondrial genes + 1 nuclear gene
Pyron et al. 2011	A,X	97-100%	761	3 mitochondrial genes + 2 nuclear genes
Gauthier et al. 2012	X,A	95-100%	141 extant + 51 extinct	610 morphological characters
Wiens et al. 2012	A,X	100%	161	44 nuclear genes
Pyron et al. 2013	X+A	95%	4,161	5 mitochondrial genes + 7 nuclear genes
Pyron et al. 2014	A,X*	99%	32	333 nuclear loci with 100% coverage
Zheng & Wiens 2016	A,X	91%	4,162	5 mitochondrial genes + 47 nuclear genes

Parafimbrios lao From Teynié et al. 2015

Within the past year, two new studies on the chromosomes of dragonsnakes (Xenodermus javanicus) have been published. In the first, the karyotype (the number of chromosomes and their shape) of dragonsnakes was reported for the first time. In humans, each cell normally contains 23 pairs of chromosomes, for a total of 46. In most snakes, each cell normally contains 18 pairs of chromosomes, for a total of 36. Usually, eight of these pairs are relatively large (called macrochromosomes), and the other ten are somewhat small (called microchromosomes). Dragonsnakes have 16 pairs of chromosomes, for a total of 32, of which seven are large and nine are small. The dragonsnake karyotype probably evolved by two fusion events, one of two macrochromosomes and the other between a macrochromosome with a microchromosome. There are some other exceptions to the 18-pair pattern; some snakes have as few as 12 or as many as 25 pairs, including the only other xenodermid to have been karyotyped, the Sichuan Odd-scaled Snake (Achalinus meiguensis), which has just 12 pairs of chromosomes.

Amami Odd-scaled Snake (Achalinus werneri)

From the 1960s to the 1980s, before DNA sequencing became cheap and easy, scientists invested heavily in collecting karyotypes from a diversity of species for comparative purposes, so we can say with pretty good certainty that the ancestral state for all snakes is 36 (18 pairs). That's the number in filesnakes, pareids, most vipers, homalopsids, and many "crown colubroids" (colubrids, lamprophiids, and elapids, although there are lots of exceptions in these three groups).  The fusions in xenodermids emphasize their uniqueness, but unfortunately don't shed any new light on their phylogenetic placement.

Stoliczkia borneensis

The other study focused on the sex chromosomes. In humans, sex is determined by which combination of sex chromosomes a baby receives from its parents: two X chromosomes make a female, whereas an X and a Y chromosome make a male. It's pretty similar in snakes, with a twist: the sex chromosomes are called Z and W instead of X any Y, and females are the heterogametic sex (meaning that a Z and a W chromosome make a female, and two Z chromosomes make a male). Birds and many other reptiles also have ZW sex determination. In many colubroid snakes, the W chromosome is about twice the size of the Z,  and it is often unusual in other ways as well, such as having sections of highly condensed chromatin or a different centromere position. In contrast, filesnakes, boids, and other more basal snakes have morphologically indistinguishable Z and W chromosomes, although they still contain different genes and perform different functions.

Perrotet's Narrow-headed Snake (Xylophis perroteti) Are members of this genus really xenodermids? Or, like the former xenodermids Oxyrhabdium and Nothopsis, will they prove to be more closely related to something else?

One reason the W chromosome looks so different from the Z in colubroids is that it contains repetitive elements called Bkm ('banded krait minorsatellite') repeats, which consist of the sequence "GATA" (sometimes "GACA") repeated thousands of times. Mammalian X chromosomes and avian W chromosomes also have these repeats. Cell biologists think that these repeats function to inactivate all the genes on the W chromosome except for those that determine sex³. Both mammalian X chromosomes and snake W chromosomes become very dense in body cells, so that none of the genes on them can be expressed. They only decondense and plays their brief, female-determining roles, in maturing eggs that are destined to become females. Unlike in mammals, the sex chromosomes of snakes span the gamut from completely identical to markedly differentiated, allowing biologists to study the evolution of chromosomal sex determination. The new study showed that female dragonsnakes have two different-looking sex chromosomes, with many Bkm repeats in the W, whereas the two Z sex chromosomes of male dragonsnakes look similar and lacked Bkm repeats, bolstering the relationship between xenodermids and other colubroids and diminishing the relationship between xenodermids and filesnakes.

The other major finding of the new study is the documentation that at least part of the sex chromosomes are homologous across all families of caenophidian snakes, suggesting that snake sex chromosomes emerged in the common ancestor of Caenophidia some 60-80 million years ago. One gene that is only on the Z chromosome in all caenophidians, including dragonsnakes, is also found on the W chromosome in filesnakes. The Z-chromosome-specific genes in caenophidians were on both the Z and W chromosomes in boas, pythons, and sunbeam snakes (Xenopeltidae), as well as in bearded dragons and anoles. Other toxicoferan lizards with ZW sex chromosomes, including chameleons and monitor lizards, seem to have evolved them independently.

ACKNOWLEDGMENTS

Thanks to Thomas Calame, Sam Howard, Konrad Mebert, Zeeshan Mirza, Takehito Sato, and Stephen Zozaya for the use of their photos.

REFERENCES